[PATCH v9 00/10] hw/ssi: imx_spi: Fix various bugs in the imx_spi model

[PATCH v9 00/10] hw/ssi: imx_spi: Fix various bugs in the imx_spi model

[Bin Meng]

From: Bin Meng <bin.meng@windriver.com>

Starting from v8, the series is based on the following 2 versions:

- v5 series sent from Bin
  http://patchwork.ozlabs.org/project/qemu-devel/list/?series=223919
- v7 series sent from Philippe
  http://patchwork.ozlabs.org/project/qemu-devel/list/?series=224612

This series fixes a bunch of bugs in current implementation of the imx
spi controller, including the following issues:

- remove imx_spi_update_irq() in imx_spi_reset()
- chip select signal was not lower down when spi controller is disabled
- round up the tx burst length to be multiple of 8
- transfer incorrect data when the burst length is larger than 32 bit
- spi controller tx and rx fifo endianness is incorrect
- remove pointless variable (s->burst_length) initialization (Philippe)
- rework imx_spi_reset() to keep CONREG register value (Philippe)
- rework imx_spi_read() to handle block disabled (Philippe)
- rework imx_spi_write() to handle block disabled (Philippe)

Tested with upstream U-Boot v2020.10 (polling mode) and VxWorks 7
(interrupt mode).

Changes in v9:
- Add a 'common_reset' function that does most of reset operation,
  leaving ECSPI_CONREG clear in imx_spi_reset().
- Do the LOG_UNIMP when the unsupported burst length value is written,
  rather than where it is used.
- Squash the 2 LOG_UNIMP warnings down into one line

Changes in v8:
- keep the controller disable logic in the ECSPI_CONREG case
  in imx_spi_write()

Changes in v7:
- remove the RFC tag
- remove the RFC tag
- remove the RFC tag
- remove the RFC tag

Changes in v6:
- new patch: [RFC] remove pointless variable initialization
- new patch: [RFC] rework imx_spi_reset() to keep CONREG register value
- new patch: [RFC] rework imx_spi_read() to handle block disabled
- new patch: [RFC] rework imx_spi_write() to handle block disabled

Changes in v5:
- rename imx_spi_hard_reset() to imx_spi_soft_reset()
- round up the burst length to be multiple of 8

Changes in v4:
- adujst the patch 2,3 order
- rename imx_spi_soft_reset() to imx_spi_hard_reset() to avoid confusion
- s/normal/common/ in the commit message
- log the burst length value in the log message

Changes in v3:
- new patch: remove imx_spi_update_irq() in imx_spi_reset()
- Move the chip selects disable out of imx_spi_reset()
- new patch: log unimplemented burst length
- Simplify the tx fifo endianness handling

Changes in v2:
- Fix the "Fixes" tag in the commit message
- Use ternary operator as Philippe suggested

Bin Meng (5):
  hw/ssi: imx_spi: Use a macro for number of chip selects supported
  hw/ssi: imx_spi: Remove imx_spi_update_irq() in imx_spi_reset()
  hw/ssi: imx_spi: Round up the burst length to be multiple of 8
  hw/ssi: imx_spi: Correct the burst length > 32 bit transfer logic
  hw/ssi: imx_spi: Correct tx and rx fifo endianness

Philippe Mathieu-Daudé (4):
  hw/ssi: imx_spi: Remove pointless variable initialization
  hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value
  hw/ssi: imx_spi: Rework imx_spi_read() to handle block disabled
  hw/ssi: imx_spi: Rework imx_spi_write() to handle block disabled

Xuzhou Cheng (1):
  hw/ssi: imx_spi: Disable chip selects when controller is disabled

 include/hw/ssi/imx_spi.h |   5 +-
 hw/ssi/imx_spi.c         | 153 +++++++++++++++++++++++++--------------
 2 files changed, 101 insertions(+), 57 deletions(-)

-- 
2.25.1

[PATCH v9 01/10] hw/ssi: imx_spi: Use a macro for number of chip selects supported

[Bin Meng]

From: Bin Meng <bin.meng@windriver.com>

Avoid using a magic number (4) everywhere for the number of chip
selects supported.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Alistair Francis <alistair.francis@wdc.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Reviewed-by: Juan Quintela <quintela@redhat.com>
---

(no changes since v1)

 include/hw/ssi/imx_spi.h | 5 ++++-
 hw/ssi/imx_spi.c         | 4 ++--
 2 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/include/hw/ssi/imx_spi.h b/include/hw/ssi/imx_spi.h
index b82b17f364..eeaf49bbac 100644
--- a/include/hw/ssi/imx_spi.h
+++ b/include/hw/ssi/imx_spi.h
@@ -77,6 +77,9 @@
 
 #define EXTRACT(value, name) extract32(value, name##_SHIFT, name##_LENGTH)
 
+/* number of chip selects supported */
+#define ECSPI_NUM_CS 4
+
 #define TYPE_IMX_SPI "imx.spi"
 OBJECT_DECLARE_SIMPLE_TYPE(IMXSPIState, IMX_SPI)
 
@@ -89,7 +92,7 @@ struct IMXSPIState {
 
     qemu_irq irq;
 
-    qemu_irq cs_lines[4];
+    qemu_irq cs_lines[ECSPI_NUM_CS];
 
     SSIBus *bus;
 
diff --git a/hw/ssi/imx_spi.c b/hw/ssi/imx_spi.c
index d8885ae454..e605049a21 100644
--- a/hw/ssi/imx_spi.c
+++ b/hw/ssi/imx_spi.c
@@ -361,7 +361,7 @@ static void imx_spi_write(void *opaque, hwaddr offset, uint64_t value,
 
             /* We are in master mode */
 
-            for (i = 0; i < 4; i++) {
+            for (i = 0; i < ECSPI_NUM_CS; i++) {
                 qemu_set_irq(s->cs_lines[i],
                              i == imx_spi_selected_channel(s) ? 0 : 1);
             }
@@ -424,7 +424,7 @@ static void imx_spi_realize(DeviceState *dev, Error **errp)
     sysbus_init_mmio(SYS_BUS_DEVICE(dev), &s->iomem);
     sysbus_init_irq(SYS_BUS_DEVICE(dev), &s->irq);
 
-    for (i = 0; i < 4; ++i) {
+    for (i = 0; i < ECSPI_NUM_CS; ++i) {
         sysbus_init_irq(SYS_BUS_DEVICE(dev), &s->cs_lines[i]);
     }
 
-- 
2.25.1

[PATCH v9 02/10] hw/ssi: imx_spi: Remove imx_spi_update_irq() in imx_spi_reset()

[Bin Meng]

From: Bin Meng <bin.meng@windriver.com>

Usually the approach is that the device on the other end of the line
is going to reset its state anyway, so there's no need to actively
signal an irq line change during the reset hook.

Move imx_spi_update_irq() out of imx_spi_reset(), to a new function
imx_spi_soft_reset() that is called when the controller is disabled.

Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

---

(no changes since v5)

Changes in v5:
- rename imx_spi_hard_reset() to imx_spi_soft_reset()

Changes in v4:
- adujst the patch 2,3 order
- rename imx_spi_soft_reset() to imx_spi_hard_reset() to avoid confusion

Changes in v3:
- new patch: remove imx_spi_update_irq() in imx_spi_reset()

 hw/ssi/imx_spi.c | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/hw/ssi/imx_spi.c b/hw/ssi/imx_spi.c
index e605049a21..4d488b159a 100644
--- a/hw/ssi/imx_spi.c
+++ b/hw/ssi/imx_spi.c
@@ -241,11 +241,16 @@ static void imx_spi_reset(DeviceState *dev)
     imx_spi_rxfifo_reset(s);
     imx_spi_txfifo_reset(s);
 
-    imx_spi_update_irq(s);
-
     s->burst_length = 0;
 }
 
+static void imx_spi_soft_reset(IMXSPIState *s)
+{
+    imx_spi_reset(DEVICE(s));
+
+    imx_spi_update_irq(s);
+}
+
 static uint64_t imx_spi_read(void *opaque, hwaddr offset, unsigned size)
 {
     uint32_t value = 0;
@@ -351,8 +356,9 @@ static void imx_spi_write(void *opaque, hwaddr offset, uint64_t value,
         s->regs[ECSPI_CONREG] = value;
 
         if (!imx_spi_is_enabled(s)) {
-            /* device is disabled, so this is a reset */
-            imx_spi_reset(DEVICE(s));
+            /* device is disabled, so this is a soft reset */
+            imx_spi_soft_reset(s);
+
             return;
         }
 
-- 
2.25.1

[PATCH v9 03/10] hw/ssi: imx_spi: Remove pointless variable initialization

[Bin Meng]

From: Philippe Mathieu-Daudé <f4bug@amsat.org>

'burst_length' is cleared in imx_spi_reset(), which is called
after imx_spi_realize(). Remove the initialization to simplify.

Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210115153049.3353008-3-f4bug@amsat.org>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Signed-off-by: Bin Meng <bin.meng@windriver.com>

---

(no changes since v7)

Changes in v7:
- remove the RFC tag

Changes in v6:
- new patch: [RFC] remove pointless variable initialization

 hw/ssi/imx_spi.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/hw/ssi/imx_spi.c b/hw/ssi/imx_spi.c
index 4d488b159a..8fb3c9b6d1 100644
--- a/hw/ssi/imx_spi.c
+++ b/hw/ssi/imx_spi.c
@@ -434,8 +434,6 @@ static void imx_spi_realize(DeviceState *dev, Error **errp)
         sysbus_init_irq(SYS_BUS_DEVICE(dev), &s->cs_lines[i]);
     }
 
-    s->burst_length = 0;
-
     fifo32_create(&s->tx_fifo, ECSPI_FIFO_SIZE);
     fifo32_create(&s->rx_fifo, ECSPI_FIFO_SIZE);
 }
-- 
2.25.1

[PATCH v9 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value

[Bin Meng]

From: Philippe Mathieu-Daudé <f4bug@amsat.org>

When the block is disabled, all registers are reset with the
exception of the ECSPI_CONREG. It is initialized to zero
when the instance is created.

Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
     chapter 21.7.3: Control Register (ECSPIx_CONREG)

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
[bmeng: add a 'common_reset' function that does most of reset operation]
Signed-off-by: Bin Meng <bin.meng@windriver.com>

---

Changes in v9:
- Add a 'common_reset' function that does most of reset operation,
  leaving ECSPI_CONREG clear in imx_spi_reset().

Changes in v7:
- remove the RFC tag

Changes in v6:
- new patch: [RFC] rework imx_spi_reset() to keep CONREG register value

 hw/ssi/imx_spi.c | 32 ++++++++++++++++++++++++--------
 1 file changed, 24 insertions(+), 8 deletions(-)

diff --git a/hw/ssi/imx_spi.c b/hw/ssi/imx_spi.c
index 8fb3c9b6d1..e85be6ae60 100644
--- a/hw/ssi/imx_spi.c
+++ b/hw/ssi/imx_spi.c
@@ -228,15 +228,23 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)
             fifo32_num_used(&s->tx_fifo), fifo32_num_used(&s->rx_fifo));
 }
 
-static void imx_spi_reset(DeviceState *dev)
+static void imx_spi_common_reset(IMXSPIState *s)
 {
-    IMXSPIState *s = IMX_SPI(dev);
-
-    DPRINTF("\n");
-
-    memset(s->regs, 0, sizeof(s->regs));
+    int i;
 
-    s->regs[ECSPI_STATREG] = 0x00000003;
+    for (i = 0; i < ARRAY_SIZE(s->regs); i++) {
+        switch (i) {
+        case ECSPI_CONREG:
+            /* CONREG is not updated on soft reset */
+            break;
+        case ECSPI_STATREG:
+            s->regs[i] = 0x00000003;
+            break;
+        default:
+            s->regs[i] = 0;
+            break;
+        }
+    }
 
     imx_spi_rxfifo_reset(s);
     imx_spi_txfifo_reset(s);
@@ -246,11 +254,19 @@ static void imx_spi_reset(DeviceState *dev)
 
 static void imx_spi_soft_reset(IMXSPIState *s)
 {
-    imx_spi_reset(DEVICE(s));
+    imx_spi_common_reset(s);
 
     imx_spi_update_irq(s);
 }
 
+static void imx_spi_reset(DeviceState *dev)
+{
+    IMXSPIState *s = IMX_SPI(dev);
+
+    imx_spi_common_reset(s);
+    s->regs[ECSPI_CONREG] = 0;
+}
+
 static uint64_t imx_spi_read(void *opaque, hwaddr offset, unsigned size)
 {
     uint32_t value = 0;
-- 
2.25.1

Re: [PATCH v9 04/10] hw/ssi: imx_spi: Rework imx_spi_reset() to keep CONREG register value

[Philippe Mathieu-Daudé]

On 1/29/21 2:23 PM, Bin Meng wrote:
> From: Philippe Mathieu-Daudé <f4bug@amsat.org>
> 
> When the block is disabled, all registers are reset with the
> exception of the ECSPI_CONREG. It is initialized to zero
> when the instance is created.
> 
> Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
>      chapter 21.7.3: Control Register (ECSPIx_CONREG)
> 
> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> [bmeng: add a 'common_reset' function that does most of reset operation]
> Signed-off-by: Bin Meng <bin.meng@windriver.com>
> 
> ---
> 
> Changes in v9:
> - Add a 'common_reset' function that does most of reset operation,
>   leaving ECSPI_CONREG clear in imx_spi_reset().
> 
> Changes in v7:
> - remove the RFC tag
> 
> Changes in v6:
> - new patch: [RFC] rework imx_spi_reset() to keep CONREG register value
> 
>  hw/ssi/imx_spi.c | 32 ++++++++++++++++++++++++--------
>  1 file changed, 24 insertions(+), 8 deletions(-)

Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>

[PATCH v9 05/10] hw/ssi: imx_spi: Rework imx_spi_read() to handle block disabled

[Bin Meng]

From: Philippe Mathieu-Daudé <f4bug@amsat.org>

When the block is disabled, it stay it is 'internal reset logic'
(internal clocks are gated off). Reading any register returns
its reset value. Only update this value if the device is enabled.

Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
     chapter 21.7.3: Control Register (ECSPIx_CONREG)

Reviewed-by: Juan Quintela <quintela@redhat.com>
Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210115153049.3353008-5-f4bug@amsat.org>
Reviewed-by: Bin Meng <bin.meng@windriver.com>
Signed-off-by: Bin Meng <bin.meng@windriver.com>

---

(no changes since v7)

Changes in v7:
- remove the RFC tag

Changes in v6:
- new patch: [RFC] rework imx_spi_read() to handle block disabled

 hw/ssi/imx_spi.c | 60 +++++++++++++++++++++++-------------------------
 1 file changed, 29 insertions(+), 31 deletions(-)

diff --git a/hw/ssi/imx_spi.c b/hw/ssi/imx_spi.c
index e85be6ae60..21e2c9dea3 100644
--- a/hw/ssi/imx_spi.c
+++ b/hw/ssi/imx_spi.c
@@ -279,42 +279,40 @@ static uint64_t imx_spi_read(void *opaque, hwaddr offset, unsigned size)
         return 0;
     }
 
-    switch (index) {
-    case ECSPI_RXDATA:
-        if (!imx_spi_is_enabled(s)) {
-            value = 0;
-        } else if (fifo32_is_empty(&s->rx_fifo)) {
-            /* value is undefined */
-            value = 0xdeadbeef;
-        } else {
-            /* read from the RX FIFO */
-            value = fifo32_pop(&s->rx_fifo);
-        }
-
-        break;
-    case ECSPI_TXDATA:
-        qemu_log_mask(LOG_GUEST_ERROR, "[%s]%s: Trying to read from TX FIFO\n",
-                      TYPE_IMX_SPI, __func__);
-
-        /* Reading from TXDATA gives 0 */
-
-        break;
-    case ECSPI_MSGDATA:
-        qemu_log_mask(LOG_GUEST_ERROR, "[%s]%s: Trying to read from MSG FIFO\n",
-                      TYPE_IMX_SPI, __func__);
+    value = s->regs[index];
+
+    if (imx_spi_is_enabled(s)) {
+        switch (index) {
+        case ECSPI_RXDATA:
+            if (fifo32_is_empty(&s->rx_fifo)) {
+                /* value is undefined */
+                value = 0xdeadbeef;
+            } else {
+                /* read from the RX FIFO */
+                value = fifo32_pop(&s->rx_fifo);
+            }
+            break;
+        case ECSPI_TXDATA:
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "[%s]%s: Trying to read from TX FIFO\n",
+                          TYPE_IMX_SPI, __func__);
 
-        /* Reading from MSGDATA gives 0 */
+            /* Reading from TXDATA gives 0 */
+            break;
+        case ECSPI_MSGDATA:
+            qemu_log_mask(LOG_GUEST_ERROR,
+                          "[%s]%s: Trying to read from MSG FIFO\n",
+                          TYPE_IMX_SPI, __func__);
+            /* Reading from MSGDATA gives 0 */
+            break;
+        default:
+            break;
+        }
 
-        break;
-    default:
-        value = s->regs[index];
-        break;
+        imx_spi_update_irq(s);
     }
-
     DPRINTF("reg[%s] => 0x%" PRIx32 "\n", imx_spi_reg_name(index), value);
 
-    imx_spi_update_irq(s);
-
     return (uint64_t)value;
 }
 
-- 
2.25.1

[PATCH v9 06/10] hw/ssi: imx_spi: Rework imx_spi_write() to handle block disabled

[Bin Meng]

From: Philippe Mathieu-Daudé <f4bug@amsat.org>

When the block is disabled, only the ECSPI_CONREG register can
be modified. Setting the EN bit enabled the device, clearing it
"disables the block and resets the internal logic with the
exception of the ECSPI_CONREG" register.

Ignore all other registers write except ECSPI_CONREG when the
block is disabled.

Ref: i.MX 6DQ Applications Processor Reference Manual (IMX6DQRM),
     chapter 21.7.3: Control Register (ECSPIx_CONREG)

Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
Message-Id: <20210115153049.3353008-6-f4bug@amsat.org>
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

---

(no changes since v8)

Changes in v8:
- keep the controller disable logic in the ECSPI_CONREG case
  in imx_spi_write()

Changes in v7:
- remove the RFC tag

Changes in v6:
- new patch: [RFC] rework imx_spi_write() to handle block disabled

 hw/ssi/imx_spi.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/hw/ssi/imx_spi.c b/hw/ssi/imx_spi.c
index 21e2c9dea3..4cfbb73e35 100644
--- a/hw/ssi/imx_spi.c
+++ b/hw/ssi/imx_spi.c
@@ -332,6 +332,14 @@ static void imx_spi_write(void *opaque, hwaddr offset, uint64_t value,
     DPRINTF("reg[%s] <= 0x%" PRIx32 "\n", imx_spi_reg_name(index),
             (uint32_t)value);
 
+    if (!imx_spi_is_enabled(s)) {
+        /* Block is disabled */
+        if (index != ECSPI_CONREG) {
+            /* Ignore access */
+            return;
+        }
+    }
+
     change_mask = s->regs[index] ^ value;
 
     switch (index) {
@@ -340,10 +348,7 @@ static void imx_spi_write(void *opaque, hwaddr offset, uint64_t value,
                       TYPE_IMX_SPI, __func__);
         break;
     case ECSPI_TXDATA:
-        if (!imx_spi_is_enabled(s)) {
-            /* Ignore writes if device is disabled */
-            break;
-        } else if (fifo32_is_full(&s->tx_fifo)) {
+        if (fifo32_is_full(&s->tx_fifo)) {
             /* Ignore writes if queue is full */
             break;
         }
-- 
2.25.1

[PATCH v9 07/10] hw/ssi: imx_spi: Disable chip selects when controller is disabled

[Bin Meng]

From: Xuzhou Cheng <xuzhou.cheng@windriver.com>

When a write to ECSPI_CONREG register to disable the SPI controller,
imx_spi_soft_reset() is called to reset the controller, but chip
select lines should have been disabled, otherwise the state machine
of any devices (e.g.: SPI flashes) connected to the SPI master is
stuck to its last state and responds incorrectly to any follow-up
commands.

Fixes: c906a3a01582 ("i.MX: Add the Freescale SPI Controller")
Signed-off-by: Xuzhou Cheng <xuzhou.cheng@windriver.com>
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>

---

(no changes since v3)

Changes in v3:
- Move the chip selects disable out of imx_spi_reset()

Changes in v2:
- Fix the "Fixes" tag in the commit message

 hw/ssi/imx_spi.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/hw/ssi/imx_spi.c b/hw/ssi/imx_spi.c
index 4cfbb73e35..2fb65498c3 100644
--- a/hw/ssi/imx_spi.c
+++ b/hw/ssi/imx_spi.c
@@ -254,9 +254,15 @@ static void imx_spi_common_reset(IMXSPIState *s)
 
 static void imx_spi_soft_reset(IMXSPIState *s)
 {
+    int i;
+
     imx_spi_common_reset(s);
 
     imx_spi_update_irq(s);
+
+    for (i = 0; i < ECSPI_NUM_CS; i++) {
+        qemu_set_irq(s->cs_lines[i], 1);
+    }
 }
 
 static void imx_spi_reset(DeviceState *dev)
-- 
2.25.1

[PATCH v9 08/10] hw/ssi: imx_spi: Round up the burst length to be multiple of 8

[Bin Meng]

From: Bin Meng <bin.meng@windriver.com>

Current implementation of the imx spi controller expects the burst
length to be multiple of 8, which is the most common use case.

In case the burst length is not what we expect, log it to give user
a chance to notice it, and round it up to be multiple of 8.

Signed-off-by: Bin Meng <bin.meng@windriver.com>

---

Changes in v9:
- Do the LOG_UNIMP when the unsupported burst length value is written,
  rather than where it is used.
- Squash the 2 LOG_UNIMP warnings down into one line

Changes in v5:
- round up the burst length to be multiple of 8

Changes in v4:
- s/normal/common/ in the commit message
- log the burst length value in the log message

Changes in v3:
- new patch: log unimplemented burst length

 hw/ssi/imx_spi.c | 17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)

diff --git a/hw/ssi/imx_spi.c b/hw/ssi/imx_spi.c
index 2fb65498c3..41fe199c9f 100644
--- a/hw/ssi/imx_spi.c
+++ b/hw/ssi/imx_spi.c
@@ -128,7 +128,14 @@ static uint8_t imx_spi_selected_channel(IMXSPIState *s)
 
 static uint32_t imx_spi_burst_length(IMXSPIState *s)
 {
-    return EXTRACT(s->regs[ECSPI_CONREG], ECSPI_CONREG_BURST_LENGTH) + 1;
+    uint32_t burst;
+
+    burst = EXTRACT(s->regs[ECSPI_CONREG], ECSPI_CONREG_BURST_LENGTH) + 1;
+    if (burst % 8) {
+        burst = ROUND_UP(burst, 8);
+    }
+
+    return burst;
 }
 
 static bool imx_spi_is_enabled(IMXSPIState *s)
@@ -328,6 +335,7 @@ static void imx_spi_write(void *opaque, hwaddr offset, uint64_t value,
     IMXSPIState *s = opaque;
     uint32_t index = offset >> 2;
     uint32_t change_mask;
+    uint32_t burst;
 
     if (index >=  ECSPI_MAX) {
         qemu_log_mask(LOG_GUEST_ERROR, "[%s]%s: Bad register at offset 0x%"
@@ -380,6 +388,13 @@ static void imx_spi_write(void *opaque, hwaddr offset, uint64_t value,
     case ECSPI_CONREG:
         s->regs[ECSPI_CONREG] = value;
 
+        burst = EXTRACT(s->regs[ECSPI_CONREG], ECSPI_CONREG_BURST_LENGTH) + 1;
+        if (burst % 8) {
+            qemu_log_mask(LOG_UNIMP,
+                          "[%s]%s: burst length %d not supported: rounding up to next multiple of 8\n",
+                          TYPE_IMX_SPI, __func__, burst);
+        }
+
         if (!imx_spi_is_enabled(s)) {
             /* device is disabled, so this is a soft reset */
             imx_spi_soft_reset(s);
-- 
2.25.1

[PATCH v9 09/10] hw/ssi: imx_spi: Correct the burst length > 32 bit transfer logic

[Bin Meng]

From: Bin Meng <bin.meng@windriver.com>

For the ECSPIx_CONREG register BURST_LENGTH field, the manual says:

0x020 A SPI burst contains the 1 LSB in first word and all 32 bits in second word.
0x021 A SPI burst contains the 2 LSB in first word and all 32 bits in second word.

Current logic uses either s->burst_length or 32, whichever smaller,
to determine how many bits it should read from the tx fifo each time.
For example, for a 48 bit burst length, current logic transfers the
first 32 bit from the first word in the tx fifo, followed by a 16
bit from the second word in the tx fifo, which is wrong. The correct
logic should be: transfer the first 16 bit from the first word in
the tx fifo, followed by a 32 bit from the second word in the tx fifo.

With this change, SPI flash can be successfully probed by U-Boot on
imx6 sabrelite board.

  => sf probe
  SF: Detected sst25vf016b with page size 256 Bytes, erase size 4 KiB, total 2 MiB

Fixes: c906a3a01582 ("i.MX: Add the Freescale SPI Controller")
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>

---

(no changes since v2)

Changes in v2:
- Use ternary operator as Philippe suggested

 hw/ssi/imx_spi.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/hw/ssi/imx_spi.c b/hw/ssi/imx_spi.c
index 41fe199c9f..a34194c1b0 100644
--- a/hw/ssi/imx_spi.c
+++ b/hw/ssi/imx_spi.c
@@ -185,7 +185,7 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)
 
         DPRINTF("data tx:0x%08x\n", tx);
 
-        tx_burst = MIN(s->burst_length, 32);
+        tx_burst = (s->burst_length % 32) ? : 32;
 
         rx = 0;
 
-- 
2.25.1

[PATCH v9 10/10] hw/ssi: imx_spi: Correct tx and rx fifo endianness

[Bin Meng]

From: Bin Meng <bin.meng@windriver.com>

The endianness of data exchange between tx and rx fifo is incorrect.
Earlier bytes are supposed to show up on MSB and later bytes on LSB,
ie: in big endian. The manual does not explicitly say this, but the
U-Boot and Linux driver codes have a swap on the data transferred
to tx fifo and from rx fifo.

With this change, U-Boot read from / write to SPI flash tests pass.

  => sf test 1ff000 1000
  SPI flash test:
  0 erase: 0 ticks, 4096000 KiB/s 32768.000 Mbps
  1 check: 3 ticks, 1333 KiB/s 10.664 Mbps
  2 write: 235 ticks, 17 KiB/s 0.136 Mbps
  3 read: 2 ticks, 2000 KiB/s 16.000 Mbps
  Test passed
  0 erase: 0 ticks, 4096000 KiB/s 32768.000 Mbps
  1 check: 3 ticks, 1333 KiB/s 10.664 Mbps
  2 write: 235 ticks, 17 KiB/s 0.136 Mbps
  3 read: 2 ticks, 2000 KiB/s 16.000 Mbps

Fixes: c906a3a01582 ("i.MX: Add the Freescale SPI Controller")
Signed-off-by: Bin Meng <bin.meng@windriver.com>
Reviewed-by: Peter Maydell <peter.maydell@linaro.org>

---

(no changes since v3)

Changes in v3:
- Simplify the tx fifo endianness handling

 hw/ssi/imx_spi.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/hw/ssi/imx_spi.c b/hw/ssi/imx_spi.c
index a34194c1b0..189423bb3a 100644
--- a/hw/ssi/imx_spi.c
+++ b/hw/ssi/imx_spi.c
@@ -169,7 +169,6 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)
 
     while (!fifo32_is_empty(&s->tx_fifo)) {
         int tx_burst = 0;
-        int index = 0;
 
         if (s->burst_length <= 0) {
             s->burst_length = imx_spi_burst_length(s);
@@ -190,7 +189,7 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)
         rx = 0;
 
         while (tx_burst > 0) {
-            uint8_t byte = tx & 0xff;
+            uint8_t byte = tx >> (tx_burst - 8);
 
             DPRINTF("writing 0x%02x\n", (uint32_t)byte);
 
@@ -199,13 +198,11 @@ static void imx_spi_flush_txfifo(IMXSPIState *s)
 
             DPRINTF("0x%02x read\n", (uint32_t)byte);
 
-            tx = tx >> 8;
-            rx |= (byte << (index * 8));
+            rx = (rx << 8) | byte;
 
             /* Remove 8 bits from the actual burst */
             tx_burst -= 8;
             s->burst_length -= 8;
-            index++;
         }
 
         DPRINTF("data rx:0x%08x\n", rx);
-- 
2.25.1

Re: [PATCH v9 00/10] hw/ssi: imx_spi: Fix various bugs in the imx_spi model

[Peter Maydell]

On Fri, 29 Jan 2021 at 13:23, Bin Meng <bmeng.cn@gmail.com> wrote:
>
> From: Bin Meng <bin.meng@windriver.com>
>
> Starting from v8, the series is based on the following 2 versions:
>
> - v5 series sent from Bin
>   http://patchwork.ozlabs.org/project/qemu-devel/list/?series=223919
> - v7 series sent from Philippe
>   http://patchwork.ozlabs.org/project/qemu-devel/list/?series=224612
>
> This series fixes a bunch of bugs in current implementation of the imx
> spi controller, including the following issues:
>
> - remove imx_spi_update_irq() in imx_spi_reset()
> - chip select signal was not lower down when spi controller is disabled
> - round up the tx burst length to be multiple of 8
> - transfer incorrect data when the burst length is larger than 32 bit
> - spi controller tx and rx fifo endianness is incorrect
> - remove pointless variable (s->burst_length) initialization (Philippe)
> - rework imx_spi_reset() to keep CONREG register value (Philippe)
> - rework imx_spi_read() to handle block disabled (Philippe)
> - rework imx_spi_write() to handle block disabled (Philippe)
>
> Tested with upstream U-Boot v2020.10 (polling mode) and VxWorks 7
> (interrupt mode).

Applied to target-arm.next, thanks.

-- PMM